EP1078332B9 - Traitement de signal video - Google Patents

Traitement de signal video Download PDF

Info

Publication number
EP1078332B9
EP1078332B9 EP00912504A EP00912504A EP1078332B9 EP 1078332 B9 EP1078332 B9 EP 1078332B9 EP 00912504 A EP00912504 A EP 00912504A EP 00912504 A EP00912504 A EP 00912504A EP 1078332 B9 EP1078332 B9 EP 1078332B9
Authority
EP
European Patent Office
Prior art keywords
video signal
motion
parts
matchable
blocks
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
EP00912504A
Other languages
German (de)
English (en)
Other versions
EP1078332A1 (fr
EP1078332B1 (fr
Inventor
Piotr Wilinski
Cornelis W. A. M. Van Overveld
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Koninklijke Philips NV
Original Assignee
Koninklijke Philips Electronics NV
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Koninklijke Philips Electronics NV filed Critical Koninklijke Philips Electronics NV
Publication of EP1078332A1 publication Critical patent/EP1078332A1/fr
Publication of EP1078332B1 publication Critical patent/EP1078332B1/fr
Application granted granted Critical
Publication of EP1078332B9 publication Critical patent/EP1078332B9/fr
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/223—Analysis of motion using block-matching
    • G06T7/231—Analysis of motion using block-matching using full search
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence

Definitions

  • the invention relates to a method and device for video signal processing. More specifically, the invention relates to motion and/or depth estimation.
  • US-A-5,793,430 discloses that motion vectors resulting from block matching are corrected for periodic structures by taking a more reliable vector from an edge of a moving object containing the periodic structure. This is done by calculating and comparing different error combinations to identify periodic structures and replacing a current motion vector with one of an adjacent pixel block either from the block to the left or from the block above, which ever yields the smaller error in the current block, or by taking a combination (e.g., mean) of both vectors.
  • a combination e.g., mean
  • the invention provides a video signal processing method and device as defined in the independent claims 1 and 4.
  • Advantageous embodiments are defined in the dependent claims.
  • the invention is based on the following recognitions.
  • the depth-estimation in video images can be based on a block-matching motion estimation approach. Some blocks that can be called non-matchable may contain not enough information for an accurate matching.
  • the invention proposes their definition and the rules to cope with them during and after the matching algorithm iterations.
  • the block-based structure from motion extraction proposes to divide the image into blocks. Before the iterative matching process the blocks are ordered. Several ways of doing it are possible. Instead of processing the blocks line by line one of the solutions is to order them following the amount of the textural information (confidence) that they contain. This way the processing starts with regions which can provide the most accurate results. The blocks containing the poorest texture and therefore leading to a possible matching on noise are processed at the last moment. During the matching process the blocks are attributed new motion values, resulting from the best matching neighbor motion. It can be considered as a motion (depth) information propagation. This approach allows the propagation of the information in the image from highly textured regions to non-textured regions.
  • the algorithm copies the information from the best matching neighbor.
  • the best matching neighbor is selected from the high-confidence neighbors we can be confident about the new motion value of the considered block.
  • the best matching block can match on noise and not on features.
  • the motion value that is copied can have a low confidence. This gives rise to the generation of erroneous motion values which are spread in a low-textured neighborhood.
  • Another problem arising during the block matching is the problem of regions containing exactly the same blocks. They may be constituted for example by repetitive texture motifs that are spatially invariant. For such blocks the best matching can be obtained for several values of motion and therefore depth. Those blocks being highly textured may become also generators of erroneous motion values.
  • non-matchable block can be a solution. It is proposed to call a block non-matchable at the moment when its matching with the neighboring blocks cannot lead to an accurate result. For such blocks it is proposed not to perform the block matching. It is proposed to call a compact region containing non-matchable blocks a non-matchable region.
  • entropy could be computed using the technique as described in [2] where only 8 angles are used, so that the computation is efficient; however, cheaper approximations to the entropy can also be used.
  • the non-matchable blocks are masked till the last iteration of the algorithm. In the last iteration the values from the blocks being in the neighborhood are copied to the region of non-matchable blocks.
  • d1 and d2 indicate motion or depth information obtained for matchable blocks.
  • N-M indicates non-matchable blocks.
  • a fat line in the upper row indicates a gradient g. This gradient g is propagated in the vertical direction, indicated by an interrupted fat line pg.
  • the value d2 is assigned to the non-matchable blocks N-M.
  • Fig. 2 shows an embodiment of a video signal processing device in accordance with the present invention.
  • An input video signal V is applied to a labeling unit L, to a motion estimator ME, and to an enhancing unit EU.
  • an output signal of the labeling unit L is applied to the motion estimator ME so as to prevent the motion estimator ME from trying to obtain motion vectors for non-matchable blocks N-M.
  • An output of the motion estimator ME is applied to a producing unit PU to obtain motion vectors for the non-matchable blocks N-M.
  • the motion vectors for all blocks of the video signal are applied to the enhancing unit to generate an enhanced video signal EV having e.g. a doubled field rate.
  • the enhanced video signal EV is applied to a display unit D.
  • a primary aspect of the invention can be summarized as follows.
  • Reliable motion / depth information can be generated for blocks having a rich texture, but not for blocks having a poor texture.
  • blocks having a rich texture there might be a problem if there is a region of identical blocks, e.g. comprising repetitive texture motifs that are spatially invariant.
  • the best matching can be obtained for several values of motion / depth. It is proposed to label both types of blocks as non-matchable blocks. These blocks are masked until the last iteration of the algorithm. Thereafter, the values obtained for the matchable blocks are copied to the neighboring non-matchable blocks.
  • the values from the neighboring matchable blocks are propagated into the non-matchable region along gradients (of texture changes) so as to avoid mixing up information from two differently moving objects.
  • the invention results in a better motion / depth estimation quality.
  • the invention is advantageously applied in depth reconstruction for static environments based on video sequences.
  • the order of the steps does not need to be the order in which the steps are recited in the claim. More specifically, the step of generating motion or depth information can be carried out before the step in which certain parts of the video signal are labeled as non-matchable; the generating step must be carried out before the labeling step if the match error is used to label parts of the video signal for which a high match error is obtained as non-matchable.
  • the labeling step is based on texture information in that little texture results in the label "non-matchable"
  • the labeling step can be carried out before the motion or depth information generating step; carrying out the labeling step before the generating step now results in the advantage that the generating step only needs to be carried out for those video signal parts that are not labeled as non-matchable.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Closed-Circuit Television Systems (AREA)
  • Television Systems (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
  • Image Analysis (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Picture Signal Circuits (AREA)

Claims (5)

  1. Procédé de traitement de signal vidéo, comprenant les étapes suivantes :
    étiqueter (L) les parties (N-M) d'un signal vidéo (V) comme parties non appariables (N-M) si une information fiable de déplacement ou de profondeur ne peut être obtenue de telles parties (N-M);
    générer (ME) une information de déplacement ou de profondeur (d1, d2) pour ledit signal vidéo (V);
    et
    produire (PU) l'information de déplacement ou de profondeur (d2) pour lesdites parties non appariables (N-M) dudit signal vidéo (V) à partir de ladite information de déplacement ou de profondeur (d1, d2) générée pour ledit signal vidéo (V);
    caractérisé en ce qu'une partie du signal vidéo est étiquetée comme non appariable si
    - l'entropie de l'information de texture de toutes les parties avoisinantes est en dessous d'un premier seuil,
    - la qualité d'appariement des parties avoisinantes est en dessous d'un second seuil, et/ou
    - les parties avoisinantes sont identiques à ladite partie du signal vidéo.
  2. Procédé selon la revendication 1, dans lequel ladite étape de production (PU) inclut l'étape de copie des valeurs des informations de déplacement ou de profondeur (d2) de parties dudit signal vidéo adjacentes auxdites parties non appariables (N-M), vers lesdites parties non appariables (N-M).
  3. Procédé selon la revendication 1, dans lequel ladite étape de production (PU) inclut l'étape de propagation de l'information de gradient (g) des parties appariables avoisinantes dudit signal vidéo (V) à l'intérieur d'une partie non appariable (N-M) de façon à prolonger les gradients (g) extérieurs auxdites parties non appariables (N-M).
  4. Dispositif de traitement du signal vidéo comprenant:
    un moyen (L) pour étiqueter les parties (N-M) d'un signal vidéo (V) comme parties non appariables (N-M) si une information de déplacement ou de profondeur fiable ne peut être obtenue pour de telles parties (N-M);
    un moyen (ME) pour générer une information de déplacement ou de profondeur pour ledit signal vidéo (V); et
    un moyen (PU) pour produire une information de déplacement ou de profondeur (d2) pour lesdites parties non appariables (N-M) dudit signal vidéo (V) à partir de ladite information de déplacement ou de profondeur (d1, d2) générée pour ledit signal vidéo (V),
    caractérisé en ce qu'une partie du signal vidéo est étiquetée comme non appariable si
    - l'entropie de l'information de texture de toutes les parties avoisinantes est en dessous d'un premier seuil,
    - la qualité d'appariement des parties avoisinantes est en dessous d'un second seuil, et/ou
    - les parties avoisinantes sont identiques à ladite partie du signal vidéo.
  5. Dispositif de traitement du signal vidéo selon la revendication 4, comprenant en outre:
    un moyen (EU) pour traiter ledit signal vidéo (V) en fonction de ladite information de déplacement ou de profondeur (d1, d2) pour obtenir un signal traité (EV), et
    un moyen (D) pour afficher ledit signal traité (EV).
EP00912504A 1999-03-18 2000-02-25 Traitement de signal video Expired - Lifetime EP1078332B9 (fr)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
EP99200849 1999-03-18
EP99200849 1999-03-18
PCT/EP2000/001556 WO2000057360A1 (fr) 1999-03-18 2000-02-25 Traitement de signal video

Publications (3)

Publication Number Publication Date
EP1078332A1 EP1078332A1 (fr) 2001-02-28
EP1078332B1 EP1078332B1 (fr) 2006-06-21
EP1078332B9 true EP1078332B9 (fr) 2006-08-30

Family

ID=8240004

Family Applications (1)

Application Number Title Priority Date Filing Date
EP00912504A Expired - Lifetime EP1078332B9 (fr) 1999-03-18 2000-02-25 Traitement de signal video

Country Status (8)

Country Link
US (1) US6563873B1 (fr)
EP (1) EP1078332B9 (fr)
JP (1) JP2002540516A (fr)
KR (1) KR20010025045A (fr)
AT (1) ATE315821T1 (fr)
DE (1) DE60028884T2 (fr)
ES (1) ES2254155T3 (fr)
WO (1) WO2000057360A1 (fr)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7447337B2 (en) * 2004-10-25 2008-11-04 Hewlett-Packard Development Company, L.P. Video content understanding through real time video motion analysis
US9418433B2 (en) * 2007-07-03 2016-08-16 Koninklijke Philips N.V. Computing a depth map
KR101682137B1 (ko) 2010-10-25 2016-12-05 삼성전자주식회사 텍스처 및 움직임 감지를 사용한 시간적으로-일관된 변이 추정을 위한 방법 및 장치
CN105306077B (zh) * 2014-06-12 2017-12-01 中国石油天然气集团公司 信号解码方法及装置

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5121202A (en) 1989-05-12 1992-06-09 Nec Corporation Adaptive interframe prediction coded video communications system
DE69422876T2 (de) * 1993-10-11 2000-06-15 Thomson Multimedia, Boulogne Verfahren und Gerät zur Korrektur von Bewegungsvektoren
KR0178231B1 (ko) 1995-08-10 1999-05-01 배순훈 계층적인 움직임 추정 기법을 이용하는 움직임 벡터 검출 방법 및 장치
US5561475A (en) 1994-12-30 1996-10-01 Daewoo Electronics Co., Ltd. Variable block matching motion estimation apparatus
US5526295A (en) * 1994-12-30 1996-06-11 Intel Corporation Efficient block comparisons for motion estimation
CN1097956C (zh) * 1995-09-06 2003-01-01 皇家菲利浦电子有限公司 分段图象序列编码的方法和系统以及解码的方法和系统
JP3929492B2 (ja) * 1995-10-25 2007-06-13 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ セグメント化画像符号化方法及びシステム並びにその復号化方法及びシステム
US5929940A (en) * 1995-10-25 1999-07-27 U.S. Philips Corporation Method and device for estimating motion between images, system for encoding segmented images
GB9623573D0 (en) * 1996-11-13 1997-01-08 Philips Electronics Nv Image segmentation
WO1998035501A2 (fr) * 1997-02-06 1998-08-13 Koninklijke Philips Electronics N.V. Procede de segmentation d'images et de poursuite d'objets, et systeme correspondant
KR100301113B1 (ko) * 1998-08-05 2001-09-06 오길록 윤곽선 추적에 의한 동영상 객체 분할 방법

Also Published As

Publication number Publication date
US6563873B1 (en) 2003-05-13
WO2000057360A1 (fr) 2000-09-28
DE60028884T2 (de) 2007-01-25
KR20010025045A (ko) 2001-03-26
ES2254155T3 (es) 2006-06-16
EP1078332A1 (fr) 2001-02-28
EP1078332B1 (fr) 2006-06-21
ATE315821T1 (de) 2006-02-15
DE60028884D1 (de) 2006-08-03
JP2002540516A (ja) 2002-11-26

Similar Documents

Publication Publication Date Title
Bergen et al. Computing two motions from three frames
JP4547066B2 (ja) ２枚のフレーム間のアフィン大域的動きの推定方法
US8149915B1 (en) Refinement of motion vectors in hierarchical motion estimation
US7039110B2 (en) Methods of and units for motion or depth estimation and image processing apparatus provided with such motion estimation unit
US20120207367A1 (en) Alignment of an ordered stack of images from a specimen
US20050249284A1 (en) Method and apparatus for generating motion vector in hierarchical motion estimation
US20060165315A1 (en) Method and apparatus for depth ordering of digital images
US7110453B1 (en) Motion or depth estimation by prioritizing candidate motion vectors according to more reliable texture information
WO2013100791A1 (fr) Procédé et appareil pour conversion ascendante de vitesse de trame hiérarchisable
US6914938B2 (en) Interlaced video motion estimation
KR20040050906A (ko) 모션 추정을 위한 장치 및 방법
US20060098886A1 (en) Efficient predictive image parameter estimation
EP1078332B9 (fr) Traitement de signal video
Tsaig et al. A region-based MRF model for unsupervised segmentation of moving objects in image sequences
US7881500B2 (en) Motion estimation with video mode detection
US20080144716A1 (en) Method For Motion Vector Determination
JP2005165969A (ja) 画像処理装置、及び方法
Patras et al. Probabilistic confidence measures for block matching motion estimation
EP1042918A1 (fr) Procede et dispositif de generation d'image statique
EP2237559A1 (fr) Halosenkung basierend auf der Hintergrundbewegungseinschätzung
CN115633178A (zh) 视频帧图像运动估计方法及相关设备
KR19990032058A (ko) 2차원 병진 움직임 벡터와 선형 회기법을 이용한 전역 움직임존재 유무 판정 방법
Smith et al. Segmentation of Multiple Motions by Edge Tracking between Two Frames.
EP1050849B1 (fr) Procédé d'estimation d'un mouvement dominant entre deux trames
JP3727731B2 (ja) 動画像の動き量推定装置

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

17P Request for examination filed

Effective date: 20010328

RBV Designated contracting states (corrected)

Designated state(s): DE ES FR GB IT

17Q First examination report despatched

Effective date: 20050315

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE ES FR GB IT

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED.

Effective date: 20060621

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

REF Corresponds to:

Ref document number: 60028884

Country of ref document: DE

Date of ref document: 20060803

Kind code of ref document: P

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: ES

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20061002

ET Fr: translation filed
PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20070322

REG Reference to a national code

Ref country code: GB

Ref legal event code: 732E

Free format text: REGISTERED BETWEEN 20090312 AND 20090318

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20090213

Year of fee payment: 10

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

Effective date: 20101029

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20100301

REG Reference to a national code

Ref country code: GB

Ref legal event code: 732E

Free format text: REGISTERED BETWEEN 20120524 AND 20120530

REG Reference to a national code

Ref country code: DE

Ref legal event code: R082

Ref document number: 60028884

Country of ref document: DE

Representative=s name: PATENTANWAELTE BRESSEL UND PARTNER, DE

REG Reference to a national code

Ref country code: DE

Ref legal event code: R082

Ref document number: 60028884

Country of ref document: DE

Representative=s name: PATENTANWAELTE BRESSEL UND PARTNER MBB, DE

Effective date: 20120626

Ref country code: DE

Ref legal event code: R081

Ref document number: 60028884

Country of ref document: DE

Owner name: FUNAI ELECTRIC CO., LTD., JP

Free format text: FORMER OWNER: IPG ELECTRONICS 503 LTD., ST. PETER PORT, GUERNSEY, GB

Effective date: 20120626

Ref country code: DE

Ref legal event code: R081

Ref document number: 60028884

Country of ref document: DE

Owner name: FUNAI ELECTRIC CO., LTD., JP

Free format text: FORMER OWNER: IPG ELECTRONICS 503 LTD., ST. PETER PORT, GB

Effective date: 20120626

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20160224

Year of fee payment: 17

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20170225

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20170225

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20180214

Year of fee payment: 19

REG Reference to a national code

Ref country code: DE

Ref legal event code: R119

Ref document number: 60028884

Country of ref document: DE

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20190903